Liquid AI
"The future of AI is local."
MIT-spinout building device-native foundation models (LFMs) that run on-device across phones, cars, robots and edge hardware.
Liquid AI, Inc. is a U.S. foundation-model company founded in 2023 as a spin-off from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) by Ramin Hasani, Mathias Lechner, Alexander Amini and Daniela Rus. The company builds Liquid Foundation Models (LFMs) - efficient, general-purpose AI models designed to run natively on processors outside of data centers, addressing the latency, privacy and hardware constraints of the physical world.
Its technology traces back to MIT research on liquid neural networks and state-space architectures. Liquid AI ships compact, hardware-aware models (including the LFM2 and LFM2.5 families, vision-language and reasoning variants, and Mixture-of-Experts models) that run on common runtimes such as llama.cpp, MLX, ONNX, CoreML, SGLang and vLLM, with some models under 1GB. Developers can fine-tune and deploy models through the Liquid Edge AI Platform (LEAP) and the Apollo tooling. For industrial and robotics use, Liquid markets memory-efficient models that run directly on edge CPUs, NPUs and GPUs to bring human-level reasoning to robots and sensors, alongside agentic orchestration across SCADA, PLC and MES systems. The company reports 40+ LFMs shipped and tens of millions of model downloads, and has announced partnerships with Mercedes-Benz, AMD, Qualcomm, Shopify and Insilico Medicine. Liquid AI operates Liquid Labs as its frontier-research arm.
Primary type & automation activities this supplier delivers:
Company Facts
Founded
2023
Primary Role
Software/Algorithm
Company Size
-
Primary Region
North America
Annual Sales
-
Funding Stage
-
Funding Total
-
Certifications & Memberships
- []
Keywords
- Liquid AI
- foundation models
- on-device AI
- edge AI
- liquid neural networks
- embodied AI
- LFM
- robotics AI
- state-space models
- Software/Algorithm
- Control & Software
- Manufacturing
Liquid AI Launches LFM2.5-2.6B Model for Local AI Applications on Small Devices
Liquid AI, a startup founded by former MIT computer scientists in 2023, has introduced LFM2.5-2.6B, a new open-weight language model tailored for agentic workloads. This model can operate entirely on local hardware, including devices as small as a Raspberry Pi, without the need for cloud inference or GPUs, making it ideal for enterprises in regulated sectors that handle sensitive data. The significance of LFM2.5-2.6B lies in its ability to support high-volume, defined tasks such as document management and workflow automation in environments with limited connectivity. The model contains 2.6 billion parameters and features a 128,000-token context window, allowing it to perform efficiently on consumer hardware while maintaining low operational costs. Looking ahead, Liquid AI's focus on edge AI applications may reshape how enterprises approach AI deployment, particularly in scenarios where latency and privacy are critical. No further timeline was disclosed at the time of publication.
Technology
Physical AI Landscape: From Digital Intelligence to the Embodied Physical World
A comprehensive industry analysis of Physical AI — covering Embodied AI, Field AI, Factory AI, VLM, World Model, VLA, and the VLA vs WAM architecture debate. Includes 20+ companies, 28 references, and the RobotToday Physical AI Landscape diagram. July 2026.
Liquid AI's smallest model yet LFM2.5-230M beats models 4X its size at data extraction, can run 'anywhere'
Liquid AI, a company founded by former MIT computer scientists, has unveiled its latest AI language model, LFM2.5-230M, which is designed for efficient data extraction and local deployment on devices such as smartphones and laptops. Released today, this 230-million-parameter model is noted for its ability to run on various hardware platforms, outperforming larger models like Alibaba's Qwen3.5 and Google's Gemma 3 in specific benchmarks. Targeting developers and engineers, LFM2.5-230M operates under a dual-use commercial license, allowing free access for individuals and companies with annual revenues below $10 million, while larger enterprises must secure a paid agreement. The model distinguishes itself by utilizing the LFM2 architecture, enabling high inference speeds with a minimal memory footprint, making it suitable for edge computing. Liquid AI's launch reflects a broader industry shift towards architectural efficiency rather than sheer parameter counts, as major AI firms focus on models with hundreds of billions of parameters. The LFM2.5-230M is specifically tailored for lightweight data extraction tasks, allowing businesses to automate processes without relying on costly cloud services. In practical applications, the model has been successfully deployed in a humanoid robot, demonstrating its capability to process complex commands efficiently. Available immediately on platforms like Hugging Face, LFM2.5-230M aims to revolutionize how enterprises manage data extraction, moving away from traditional, rigid systems to more adaptable AI-driven solutions.
Technology
Laboratory Robotics in 2026: Technology, Companies & Vertical Maturity
From clinical diagnostics to self-driving labs: a deep analysis of laboratory robotics technologies, mass-production maturity timelines across 12 vertical sectors, 26 key companies, and the investment landscape shaping lab automation through 2030.
AGRICULTURAL ROBOTICS | $34B Weeding Robot Market
Weeding & crop management robots target a $34–42B herbicide market. Deep-dive on Carbon Robotics, Ecorobotix, Verdant, FarmWise, and the four tech platforms reshaping chemical-free farming by 2030.
Space Robot- Building Skyscrapers in Space: ISMA's Industrial Revolution
ISAM technology breaks rocket fairing limits through robotic space manufacturing. Market grows from $3.5B to $14.5B by 2032. See how 3D printing, autonomous assembly, and orbital fabrication are creating off-Earth factories.
Space Robotics: The 2026 Breakout
Space robotics transforms in 2026 as Starship cuts launch costs 96%, AI enables autonomy, and satellite servicing demand explodes. The breakout year explained.
Robotics Innovator Lens Unveils Ambitious Production Goals Amid AI Surge
LENS sets bold production targets as AI-augmented robotics demand soars, promising high-volume manufacturing and next-gen automation tech.
Revolutionizing Robotics: LCE-Integrated Soft Skin for Enhanced Medical Applications
Discover a new class of millimeter-scale, fully soft everting robots featuring LCE-integrated skins for high-dexterity steering. This Science Advances study details how Liquid Crystal Elastomer actuators enable large bending angles and multi-point control for advanced surgical and industrial navigation.
